PROJECT SUMMARY DESCRIPTION: Dementia, especially Alzheimer's dementia (AD), is a growing public health problem with a prevalence of 5M in the US alone (33M worldwide). Despite a decrease in incidence rates, with the aging of the population, the prevalence of dementia is expected to increase to 16M in the US (115M worldwide) with associated costs rising to $1T. Delaying long-term care by 1 month for older Americans would save $60B annually in direct care cost. Efforts to prevent or delay dementia have been largely unsuccessful. However, major depressive disorder in late life (?late-life depression?, LLD) has been identified as one of six treatable risk factors for dementia, especially AD and vascular dementia. The depression-dementia relationship may be magnified in elders who do not respond to antidepressant treatment and experience persistent symptoms. Thus, resolving whether those with treatment-resistant late-life depression (TRLLD) are at higher risk of cognitive decline and progression to dementia compared to those with treatment-responsive LLD is critically important. Leveraging a Patient-Centered Outcomes Research Institute (PCORI)-funded treatment study of N=1500 people with LLD, across 5 sites, we propose to comprehensively delineate neurocognitive and neuroimaging biomarkers associated with progression to dementia in people with persistent LLD (i.e., TRLLD) compared to those whose LLD remits with treatment. We anticipate enrolling 750 elders with LLD and characterizing their symptomatic trajectory over 24 months. We will assess each participant at three time points with neurocognitive and advanced neuroimaging. We hypothesize that changes in executive functions and the executive control network, as well as changes in episodic memory and the default mode/cortico-limbic network, will be greater in those with TRLLD than in those who respond to treatment and stay well. We also hypothesize that changes over two years in executive function and episodic memory will be specifically associated with changes in executive- control and cortico-limbic circuitry, respectively. Based on our recent findings that inflammatory and related molecular markers can differentiate those with neurocognitive impairment and LLD from those with LLD alone, we will build a predictive multivariate model combining baseline neurocognitive, neuroimaging, and plasma protein data to determine who is at greatest risk for cognitive decline and dementia. Finally, we will also explore whether latent class trajectories of depressive symptoms can go beyond the dichotomy of remission/non-remission to identify subsets of elders with LLD at highest risk of cognitive decline, neural circuit change, and progression to dementia. This work will set the stage for neural circuit- targeted preventive care to delay dementia in subsets of older patients with LLD. If successful, our work can accelerate therapeutic efforts and innovation targeting the depression- dementia pathway and reduce suffering for large numbers of elders and their families.